On Thu, Dec 19, 2013 at 12:19 PM, Wood, Peter (ISS) <p.w...@lancaster.ac.uk> wrote: > Hey, > > Given what I can see on our primary collector running ext4 (2Tb/134 million > inodes) it appears it's assigned 64 inodes per 1mb of storage available. If > you're similar I'm guessing your average size per cap files is < 16kb if > you've run out of inodes before space. > > You can override inode count when the file system is formatted (mke2fs -N), > this may come with other dangers though, I'm not a filesystem expert. If you > can shift your data somewhere else temporarily, reformat and restore this > might be an option for you.
ok. > If you can live with managing older flows outside nfsen (kinda defeats the > point though), and you don't need instant access to the older captures > regularly (i.e you're just keeping it for compliance) you could just tar them > up and delete the individual ones off disk. yes, we want nfsen for history reasons as well, to be able to see what has happened. Maybe not for months/years, but definately for weeks. > Otherwise, you could merge devices into fewer (same collector for many > devices), you'll lose the ability to reference them individually which may be > a huge issue for you. Depends on your use case, per layer collector may not > be a bad option, for example we run a Boarder, Core, Aggregation and Edge > collectors. Given we have over 1k edge devices which can sflow, an individual > profile per device would be crazy (and a royal PIA for adhoc monitoring). ok, I have not looked into this, we have a 80 node wan, so nothing in that scale. > I'm guessing the least preferable choice is to reduce your retention period > to fit within your inodes, 1 file/5 min/device. Remember to leave a few for > directory structure and any other misc FS data (does meta data consume > inodes?). this is a 16Gb ext4 disk, with lvm. Som info on the filesystem: $ sudo tune2fs -l /dev/vg_host/lv_root | grep -i inode Filesystem features: has_journal ext_attr resize_inode dir_index filetype needs_recovery extent flex_bg sparse_super large_file huge_file uninit_bg dir_nlink extra_isize Inode count: 950976 Free inodes: 183033 Inodes per group: 8128 Inode blocks per group: 508 First inode: 11 Inode size: 256 Journal inode: 8 Thanks for your insights. -- groet, natxo ------------------------------------------------------------------------------ Rapidly troubleshoot problems before they affect your business. Most IT organizations don't have a clear picture of how application performance affects their revenue. With AppDynamics, you get 100% visibility into your Java,.NET, & PHP application. Start your 15-day FREE TRIAL of AppDynamics Pro! http://pubads.g.doubleclick.net/gampad/clk?id=84349831&iu=/4140/ostg.clktrk _______________________________________________ Nfsen-discuss mailing list Nfsen-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nfsen-discuss
